# Reading Responses 5/5 Done (Set 2)
## Mar 24 Tue - Manipulated
When shopping on online stores like Amazon, I personally always check the reviews to see what other people have received vs what is advertised. Professor Reagle describes how these online reviews may actually be manipulated for commercial gain. This is done through fake positive reviews or even suppressing or deleting negative ones. This has happened to me when a restaurant in Boston used to give away free figurines if you gave them a five-star review.

Professor Reagle mentions “ice cube” framing, where the content is as featureless as it can get, and relates it to how fake these reviews can be. According to Fowler, these reviews can even be illegal. The FTC has been fining businesses for buying fake reviews. This is due to them having real consumer harm and affecting the people buying these products. I am curious as to what we may see in the future that will fix this widespread issue.
## Apr 03 Fri - Artificial intelligence
Did you know that every face you see in a dream is one you have actually seen in real life? Well, this is exactly what artificial intelligence does when it generates an image. These generative models take a variety of different photos and merge them together, consequently creating one image out of many others instead of creating an entirely new one. Newhauser described two different types of these models: transformers and Diffusion models. A transformer model runs based on large text databases, whereas diffusion models power image generation partially through fine-tuning. This concept of fine-tuning works by building off of other generative artificial intelligence products while also using the parts of past models that operated the best. Although I feel that artificial intelligence is an ethical way for less fortunate students to gain access to new information, I do feel that the generative video and image aspects are being overused, causing long-term harm to the Earth.

While Newhauser described how artificial intelligence works, Gold compares and contrasts different ways AI behaves. Sydney, being one of Microsoft’s earliest AI chatbots, performed very poorly. This performance was due to the bot re-sending an entire conversation history as the user typed a new message. As the chatbot got overloaded, it started threatening and proclaiming its love to its users. I am curious as to how this changed because now AI can remember a conversation I had with it many months ago if I ask it to remember a question. On the other hand, Spotify’s DJ works without any words. They coded this DJ to only play songs it thinks the user will like, with no room for error, because the user cannot communicate with it, only skip and play songs. Gold also described Speedy, a fictional robot, to argue how AI needs rules that humans can understand, which is something Sydney lacked. agree with this argument as I think AI should be predictable and not even be coded to be able to threaten a user for any reason.

## Apr 07 Tue - Algorithmic bias
A common algorithm that ranks colleges is actually extremely biased. This is due to this algorithm thinking it is just doing mathematical equations to rank based on data, but the data was shaped by a bias with the algorithm then amplifying it. O’Neil says this idea is what makes rankings such as baseball and colleges so biased. When someone creates a system with a biased mindset, it causes that system to also become biased while presenting itself as being based only on data, but that simply is not true. A bias in the algorithm could then lead to harm, with an example being college tuition at these high-ranking colleges increasing due to the placement given against other schools.

Rutherform and White argue that racial biases surface due to the platform’s algorithm amplifying this content. I agree with this due to the research that I have been doing for my essay. I have found that many algorithms ignore hate vs heartwarming content and promote what gets the platform the most views. They go into detail about how searching different words can lead to racist images, with context that makes it clear what the image is trying to convey to its audience. I decided to look up one of the phrases, “black hands,” and this is what came up:
As seen in the photo, nothing racist came up. This may be due to my browser history, the browser I am using, or even due to this article being outdated, as it is from 2016.

Hochman’s article can be related to O’Neil’s as they both argue how the creator of a site or algorithm can cause bias within the system that can be undetected to outside viewers. Hochman argues how ChatGPT has a built-in left-wing ideology that the AI was trained to have. Hochman used a drag queen example where the AI would only write about a drag queen in a positive light and not a negative one, which leads back to both being left-leaning, as it refused to be negative towards a drag queen. I am unsure as to if ChatGPT can be called left-leaning based on this one drag queen example, as I would code my AI to not talk negatively about anyone, but I can understand how these biases can become prevalent based on the creator. I have noticed a difference between asking ChatGPT for advice and asking Claude. Where Claude will be honest and let you know an unbiased opinion on the situation, ChatGPT will agree with your opinion, saying you can do no wrong and not disagree with what you need help with.
## 14 Tue - Digital language and generations
Have you ever needed to put your age to view a website or to use an app? Well, that data is only so it knows what content it is allowed to show you, not what content it will show you. McCulloch argues that internet language is more based on when a user first went online, rather than their actual age. I find this really interesting because I used to always wonder why so many apps requested my age, only to find out it was more asking what content I should be allowed to see, and not what content it would be recommending to show me. McCulloch used the common term “lol” as a way to describe the similarities and differences between each group of internet users. While the older generation used this phrase as actual laughter, the newer generations on the internet use this phrase as more of a sign of amusement.

McCholluch dives deeper into this concept in an interview where lol no longer has its literal meaning to the newer generation. As she stated before, “lol” has a new meaning that goes deeper than being a sign of amusement. This new meaning uses “lol” as a form of softening a once-rude message, which then leads to many miscommunications. Another example of newer internet language is “key shmashing.” This is a description of when an emotion is so overwhelming that it is incoherent, so the user matches it with text, but this has gotten so out of hand that people will delete and edit their key smashing until they are satisfied with it. I feel like both the passage and the interview are a way of McColluch telling her readers not to overthink a simple period or a different use of slang.
## Apr 17 Fri - Pushback
I wish there were a statistic on students who attended a school with a phone ban vs. a school without that restriction, and their phone usage. From what I have noticed, as someone who went to private school with no phone, I have noticed that, especially during class, some students cannot seem to live without checking their phones. Morrison and Gomez mention this in their article and argue that people who have constant internet access are always online. This idea has then led to another idea called “pushback,” where these people will limit their phone or app usage altogether. I find this concept interesting because I have seen it in some friends, but I wonder what the age range is of people becoming more aware of their phone usage, and when they choose to limit their usage.

Vadkul’s article follows a group of teens in college who are following the idea of “pushback,” where they intentionally limit their usage of all technology. I personally can relate to this because other than doing schoolwork, I find myself with my friends and rarely touching my phone, as I incredibly enjoy living in the moment. These students switched to flip phones and met in person rather than talking online to create meaningful connections. This article not only mentions the advantages of this lifestyle but also the hardships it brings to these students. College requires phones to be accessible for day-to-day lives, such as QR codes and technology, such as Canva, which limits removing internet access fully. I enjoyed these articles and felt connected to them more than others I have read in this class. I thoroughly enjoy not being on my phone a lot and would much rather see people in person to form more personal connections.